Methodological considerations for the use of faecal nitrogen to assess diet quality in ungulates: The Alpine ibex as a case study

Faecal indices are used to describe spatiotemporal patterns of diet quality in wild herbivores, and near infrared spectroscopy methods (NIRS) distinctively reduce costs and labour compared to conventional chemical analyses. In this study we compared the prediction accuracy of faecal nitrogen (N), ash and fibrous fractions of a la- boratory instrument (NIRS) with that of a lightweight portable instrument (VisNIRS) using faecal samples of Alpine ibex (Capra ibex). We then compared how different indices based on faecal N and fibrous fractions were able to detect the expected temporal trend and difference between sexes in diet quality, and verified whether the presumably lower accuracy of NIRS and VisNIRS (compared to wet chemistry) influenced the ability of indices to detect such trends. The laboratory NIRS was more accurate than the portable VisNIRS, especially for fibrous fractions. Predicted data from both instruments clearly identified the temporal trend identified by chemically analysed data, but only NIRS identified the small difference between sexes. Expressing faecal N as a proportion of organic matter (OM) or neutral detergent fibre (NDF) improved the power of the index as compared to expressing it as a proportion of dry matter (DM), irrespective of the method used to estimate the data. Faecal NDF contents followed the patterns expected from the variation of faecal N, while faecal acid detergent fibre (ADF) and especially lignin (ADL) followed contradictory patterns, which might be due to a selection of different forage sources rather than to changing quality of a homogenous forage source. We conclude that expressing faecal N as a proportion of OM is advised with wild herbivores faeces, where ash content can be biased by accidental soil contamination or ingestion. Faecal fibrous fractions are useful to complement faecal N and to avoid simplified interpretations of the results. For this purpose, NIRS methods can be extremely useful to predict multiple indices on large numbers of samples. VisNIRS portable instruments might be unable to detect subtle biological patterns in small numbers of samples. However, developing procedures for portable instruments outside laboratory facilities might further improve practicability, cost effectiveness, and, through allowing on- field predictions on large numbers of samples, power of statistical testing of diet quality indices